Biosensors derived from fluorescent proteins can now be manipulated to respond to specific molecular events within living cells. However, the application of resonance energy transfer (RET) based protein biosensors to a wide rage of medically important problems is currently limited by the lack of suitable donor and acceptor partners for efficient RET. New combinatorial approaches for generating molecular diversity wilt be coupled with quantitative cell sorting instrumentation in order to purposefully manipulate and design broadly applicable biosensors and component proteins. This approach will be used to create biosensors for use in the isolation and development of therapeutics, for improved medical diagnostic tools, and to enable real-time studies of protein-protein interactions and protease activities in living cells. [unreadable] [unreadable] Intracellular biosensors consisting of two fluorescent proteins capable of resonance energy transfer (RET) and joined by a peptide linker will be constructed and optimized. An expression system will be developed in bacteria which allows high-throughput screening of fluorescent protein variants directly for their ability to undergo RET using flow cytometry. Large designed protein libraries will be constructed in bacteria, containing more than ten million variants in which the chromophore and surrounding residues are randomized. Multi-parameter cell sorting will be Used to quantitatively screen for several criteria important for RET, allowing identification of rare; improved partners. The spectral properties of improved variants will be further characterized by fluorimetry, and a comprehensive database of sequence-function relationships will be developed using data from both previous and current studies. The sensitivity of detecting RET using flow cytometry wilt be improved using a flexible optical bench design that allows optimization of light sources and optics for biosensor excitation and detection. Methods will be developed for ultra high-throughput screening (>100,000/s) of large libraries. Biosensor expression levels, total signal-to-noise ratios, false positive events, and sort gating will be optimized for FCM screening on the basis of RET biosensor signals.